American Journal of Respiratory and Critical Care Medicine

Hypoxia-induced lymphocyte dysfunction may be implicated in endothelial cell damage in obstructive sleep apnea (OSA) syndrome. γδ T cells' unique migration, cytotoxic features, and accumulation in atherosclerotic plaques are considered critical in cardiovascular disorders. We characterized the phenotype, cytokine profile, adhesion properties, and cytotoxicity of γδ T cells in patients with OSA and control subjects. The following is a summary of our major findings regarding OSA γδ T cells: (1) a significant increase in the expression of the inhibitory natural killer B1 receptors was found; (2) the intracellular content of proinflammatory cytokines tumor necrosis factor (TNF)-α and interleukin-8 was increased, and the content of the antiinflammatory cytokine interleukin-10 was decreased; (3) γδ T cells of patients with OSA adhered significantly more avidly to nonactivated endothelial cells in culture than those of control subjects; (4) L-selectin expression was higher; (5) anti–E/P-selectin antibodies and anti–TNF-α antibodies decreased the adhesion index of OSA γδ T lymphocytes/endothelial cells but not of control subjects; and (6) cytotoxicity of OSA γδ T lymphocytes against endothelial cells in culture was 2.5-fold higher than that of control subjects and could be prevented by pretreatment with anti–TNF-α. Collectively these data implicate γδ T lymphocyte function in atherogenic sequelae in OSA.

Obstructive sleep apnea (OSA) syndrome is closely associated with cardiovascular morbidity, particularly with hypertension (13). Sleep apnea–related oxidative stress and a state of inflammatory cell activation resulting in endothelial cell injury have been proposed as one of the underlying mechanisms responsible for atherogenesis in this syndrome (4, 5).

A recent study from our laboratory has demonstrated hypoxia-related increased monocyte activation in patients with OSA (4). This was manifested by increased expression of adhesion molecules and increased production of reactive oxygen species by OSA monocytes. Moreover, adhesion of OSA monocytes to endothelial cells in culture was significantly enhanced, all implicating altered monocyte functions in the pathogenesis of OSA (4). On the basis of these findings we postulated that hypoxemia and possibly hypercapnia experienced nightly by these patients could induce activation in other circulating inflammatory cells, namely lymphocytes. Cytokineactivated T lymphocytes were shown to mediate endothelial cell activation and to further release proinflammatory cytokines as interleukin (IL)-6, IL-8, and monocyte chemotactic protein-1 in a contact-dependent manner and thereby to induce endothelial cell damage via leukocyte infiltration (6). Increased plasma levels of tumor necrosis factor (TNF)-α, IL-6, IL-8, and monocyte chemotactic protein-1 were reported in OSA (79), thus, implying that both lymphocytes and endothelium are activated.

There are two main distinct lineages of T cells, those expressing αβ T cell receptors and those expressing γδ T cell receptors (10, 11). γδ T cells differ from αβ T cells regarding tissue localization, antigen recognition, and role in inflammatory processes and tissue repair (11, 12). In humans, only a small proportion (< 5%) of circulating lymphocytes consists of γδ T cells. However, they are more widespread within epithelial-rich tissues, such as the skin and the respiratory, digestive, and reproductive tracts, where they can comprise up to 50% of T cells. This, and the fact that γδ T cells recirculate continually by migrating selectively into tissues, and back to the blood (13), account for their large heterogeneity (10, 14). Importantly, T lymphocytes, like macrophages, were shown to infiltrate the intima of arteries during the initial stages of atherosclerosis (15, 16), most of which are activated, and a high percentage (10–15%) express γδ T cell receptors (17, 18). In addition, γδ T cells have the capacity to recognize the main proatherogenic antigens such as complex lipids and heat shock proteins and exhibit higher transmigratory abilities through endothelial cells than αβ CD4 and CD8 T lymphocytes (18, 19). Also, in contrast to αβ T cells, they display major histocompatibility complex–unrestricted cytotoxicity against endothelial cells (20). Of note, activated γδ T cells can induce endothelial permeability by cytolytic processes involving the participation of adhesion pathways via γδ T cells/endothelial cell interactions (21). Activated γδ T cells were also shown in diabetes mellitus (22), arthritis (23), and scleroderma where endothelial cell damage is a prominent mechanism (20, 24). Taken together, these findings led us to assume that γδ T cells can contribute to endothelial cell damage in OSA as well.

The general objective of this study was to assess the potential involvement of circulating γδ T lymphocytes in endothelial cell damage in OSA. Specifically, we determined γδ T cell phenotype, cytokine profiles, and interactions with human umbilical vein endothelial cells (HUVECs) in culture. These were investigated with a particular emphasis on adhesion and cytotoxicity.

Patients and Control Subjects

Blood was withdrawn after overnight fasting from a total of 34 patients with OSA and 19 control subjects without apnea. Both patients and control subjects were recruited from the patients' population of the Technion Sleep Medicine Center. Patients inclusion criteria were age—30 to 75 years, a sleep laboratory finding of respiratory disturbance index (RDI) more than 15, and not being sick during the last week before the study. Control subjects without apnea were recruited from patients who were referred for polysomnographic monitoring because of suspected sleep apnea but were found to have primarily snoring or had a RDI less than 15, provided that most of the disordered breathing events were hypopneas not associated with arterial oxygen desaturation. The reason for using a cutoff threshold of RDI less than 15 rather than the customary RDI less than 5 or RDI less than 10 is attributable to the fact that since the introduction of the pressure canula to measure airflow in our laboratory, very few studies have resulted in RDI less 10, including in normal individuals. The clinical, demographic, and biochemical data of the two groups demonstrated that patients with OSA had a moderate to severe syndrome with a mean RDI of 33.1 ± 13.9 events/hour (range 18–65) and percent time below 90% SaO2 16.0 ± 24.6%. Seventeen of the 19 control subjects were recorded in the sleep laboratory (mean RDI = 9.4 ± 5.4, percent time below 90% SaO2 1.9 ± 3.4%). In the other two control subjects, OSA was ruled out by a detailed clinical interview. There were no significant differences in age (53.5 ± 12.9 vs. 53.2 ± 10.4 years), body mass index (29.36 ± 5.5 vs. 27.3 ± 3.4 kg/m2), sex distribution (76.5 vs. 77.8% males), current smokers (20.7 vs. 28.6%), rate of major diseases, or usage of medications between patients with OSA and control subjects, respectively (see online supplement for more details on demographic, clinical, and biochemical data). Because the tests could not be performed simultaneously on all subjects and the design was based on investigating OSA and control cellular functions at the same time for a given assay, different subgroups of patients and control subjects participated in each of the tests. The number of subjects who participated in each experiment is indicated. The protocol was approved by the local Human Rights Committee, and all participants signed an informed consent form.

Isolation of γδ T Cells

Enriched γδ T cells were prepared by magnet separation, as described previously (25). In all cases, the γδ T cells were 85 to 90% pure, as assessed by flow cytometry analysis, and with cell viability up to 90% as determined by trypan blue exclusion. In addition, enriched γδ T lymphocytes were prepared using the RosetteSep antibody cocktail directed against CD16, CD19, CD36, CD56, T cell receptors αβ, and glycoprotein A (StemCell Technologies Inc, Vancouver, BC, Canada).

Flow Cytometry

The cell phenotype was assessed in whole blood lymphocytes by flow cytometry (FACS Calibur; Becton Dickinson, Lincoln Park, NJ), using a single or a dual-staining protocol. The percentage of fluorescent cells and that of mean fluorescence intensity were determined in each case. Cell viability was determined by propidium iodide staining (26).

Intracellular Cytokines

Intracellular TNF-α, IL-8, IFN-γ, and IL-10 were detected in γδ T cells by using a Cytofix/Cytoperm Kit (PharMingen, San Diego, CA). The results are expressed as the percentage of γδ T cells, which contain each of the cytokines investigated. All procedures were performed according to the manufacturer's instructions.

Endothelial Cell Culture

HUVECs were kindly provided by Dr. N. Lanir (Rambam Medical Center, Haifa, Israel) and treated as described previously (4). Briefly, HUVECs were grown in 199 medium supplemented with 15% fetal calf serum (Biological Industries, Kibutz Beth HaEmek, Israel) and endothelial mitogen. After detachment with trypsin–ethylenediaminetetraacetic acid, HUVECs were seeded onto fibronectin pretreated (50 μl/well at 10 μg/ml) 96-well microplates. The HUVECs were employed only in the first and second passages.

Adhesion Assay

This assay determines the ability of lymphocytes to adhere to endothelial cells in culture by using radiolabeled lymphocytes. The 51Cr-labeled purified γδ+ or γδ− T cells were added to HUVEC monolayers for 45 to 60 minutes at 37°C at a ratio 3:1. Radioactivity of the adherent cells was determined in triplicates. The percentage of adhering cells was calculated as described previously (4). In some experiments, adhesion assays were performed in the presence of anti-human TNF-α (500 ng/ml) or using HUVECs pretreated with anti-CD62E/CD62P antibodies (against P- and E-selectin).

These experiments were also verified by fluorescence microscopy. Briefly, peripheral blood mononuclear cells depleted of plastic adherent cells (monocytes) were stained with fluorescein isothiocyanate–conjugated monoclonal antibodies against γδ T cells and then added to HUVEC monolayers for 45 minutes at 37°C (10,000 cells/well). After washing, adherent cells were quantified microscopically.

Cytotoxicity Assay

This assay determines the ability of lymphocytes to kill endothelial cells in culture by employing radiolabeled endothelial cells. HUVECs were loaded overnight with 51Cr (1 μCi/ml). After washing, purified γδ T cells were added to the 51Cr labeled HUVECs (ratio 5:1). After 2 days of coculture, counts-per-minute radioactivity was determined and calculated, as described previously (25). In order to study the effects of TNF-α on HUVECs/γδ T cells interactions, in 6/14 experiments, cytotoxicity assays were also performed in the presence of TNF-α neutralizing antibodies (500 ng/ml of recombinant anti-human TNF-α, clone B-C7; BioSource International, Inc., Cammarillo, CA).

DNA fragmentation was determined by propidium iodide labeling (26).

Statistical Analysis

Data are expressed as mean ± SD. Differences between OSA and control groups were first evaluated by a Wilcoxon test for nonparametric variables or by a t test for independent groups for parametric variables. Then, analysis of the covariance was used to compare the results of the immunologic testing using triglycerides, which was significantly different between patients and control subjects (see online supplement) and body mass index, as covariates. Nonadjusted and adjusted means of all variables are presented in the online supplement.

Immunophenotyping of γδ T Lymphocytes

We studied the percentage of peripheral blood γδ T cells in 28 out of 34 patients and in 12 out of 19 control subjects. Normally, the percentage of circulating γδ T cells in whole blood ranges between 2 and 5%. Analysis of their distribution confirmed these values for both study groups. In patients with OSA, the average percentage was 4.5 ± 3.3% and in controls without apnea 2.5 ± 3.4%, (p = 0.1). The range, however, was largely variable in patients with OSA (0.6–18%) as opposed to that in control subjects (1.5–4.0%). Similarly, the intensity of expression as attested by mean fluorescence intensity did not vary significantly between patients and control subjects (124 ± 29 and 113 ± 26, respectively; p = 0.39).

Because there is a close association between cytotoxicity and expression of NK receptors, we examined the distribution of NKB1, CD56, and CD16 on γδ T cells. The percentage of cells expressing NKB1 molecules on γδ T cells of patients with OSA (n = 16) was significantly higher than that of control subjects (n = 9) (8.2 ± 5.2 vs. 1.4 ± 5.3, p = 0.006). The individual data for both study groups are presented in Figure 1

. As can be seen, smoking, which potentially could affect the results, could not account for the differences between groups, although NKB1 values were relatively high in two control smokers.

Unlike NKB1, no differences were detected between patients (n = 17) and control subjects (n = 9) in the percentage of γδ T cells expressing CD56 receptors (double-staining CD56+/γδ+). These values for patients with OSA and control subjects were 55.9 ± 17.4% (range 30–83%) and 42.0 ± 17.8%, (range 17–84%), respectively (p = 0.08). Double-staining γδ T cells for CD16 populations (CD16+/γδ+) also revealed no differences between patients (n = 15) and control subjects (n = 6) (data in online supplement).

Cytokine Profile of γδ T Cells

The percentage of γδ T cells expressing intracellular content of the cytokines TNF-α, IL-8, IL-10, and IFN-γ was determined in freshly isolated nonstimulated γδ T cells. This presumably reflects the in vivo induction of cytokine synthesis and turnover. As depicted in Table 1

TABLE 1. The percentage of γδ t cells of patients with obstructive sleep apnea and control subjects containing different cytokines

Percentage of γδ T Cells Expressing Cytokines

Patients with OSA (n)
Control Subjects (n)
p Value
TNF-α17.5 ± 4.9 (11)4.4 ± 4.9 (9)0.0001
IL-812.9 ± 3.7 (9)6.4 ± 3.8 (7)0.009
6.1 ± 4.8 (8)
14.7 ± 4.9 (6)

Definition of abbreviations: IL = interleukin; OSA = obstructive sleep apnea; TNF = tumor necrosis factor.

Intracellular TNF-α, IL-8, and IL-10 were detected by flow cytometry using a dual-staining protocol. Adjusted means and SDs are presented, and p values for patients with OSA vs. control subjects.

, a fourfold and twofold increase in the percentage of γδ T cells of patients with OSA containing intracellular TNF-α and IL-8 were observed as compared with that of control subjects (p = 0.0001 and 0.009, respectively). By contrast, the percentage of intracellular IFN-γ containing cells varied greatly and did not differ between groups (7.5 ± 5.3% and 9.1 ± 5.4%, respectively; p = 0.7). Because TNF-α, IL-8, and IFN-γ are known for their proinflammatory effects, we also studied IL-10 which is an antiinflammatory cytokine. Thus, the percentage of OSA γδ T cells containing IL-10 was lowered to 41.5% of control values (p < 0.01). Furthermore, the percentage of OSA γδ T cells containing IL-10 was inversely correlated with the percentage of TNF-α–containing cells as illustrated in Figure 2 (r = −0.6, p < 0.05).

Lymphocyte Binding to HUVECs

In order to assess the abilities of lymphocytes from patients with OSA and control subjects to adhere to HUVECs in culture, peripheral blood mononuclear cells were depleted of monocytes (by plastic adherence) to obtain only lymphocytes. A total of 10,000 lymphocytes were stained with fluorescein isothiocyanate– conjugated monoclonal antibodies against γδ T cell receptors and cocultured with HUVEC monolayers for 45 minutes. After intensive washing, only a few cells (2–4 cells/well) from control subjects (n = 4) adhered to HUVECs, whereas from patients with OSA(n = 6), the number was substantially higher (17–21 cells/well). Fluorescence microscopic evaluation revealed that the high adhesion response of T lymphocytes from patients with OSA was largely attributed to the γδ T cell subsets. To confirm this observation, we determined the ability of purified γδ+ and γδ− T cells to adhere to HUVECs in culture as well. As expected, the ability of γδ− T cells (which include CD4+ and CD8+ T cells) from patients with OSA to adhere to nonactivated HUVECs under the same experimental conditions was low and did not differ from that of control subjects (1.8 ± 0.4, n = 5, vs. 1.7 ± 0.4, n = 4, respectively). On the other hand, the adhesion index for γδ+ T cells of patients with OSA was more than twofold higher than that for control subjects (6.1 ± 1.1 vs. 3.0 ± 0.7, p < 0.001). Figure 3

displays the individual data of adhesion indices for γδ+ T cells studied. As can be seen smoking did not alter the results in either of the study groups.

Involvement of TNF-α, L-Selectin, and NK Receptors in Lymphocyte Binding to HUVECs

The relative contributions of TNF-α, L-selectin, and NK receptors to the adhesion of OSA and control γδ T cells/HUVECs in culture were assessed as well. The involvement of TNF-α in the adhesion response was investigated by employing antibodies neutralizing TNF-α. Its addition to OSA γδ T cells/HUVECs cocultures, lowered the adhesion index by 52.5% ( p = 0.0002). By contrast, adding anti–TNF-α to control γδ T cells/HUVECs cocultures did not affect the adhesion index in 6/6 subjects (Table 2)

TABLE 2. The involvement of tumor necrosis factor-α and l-selectins in the adhesion of γδ t cells from patients with obstructive sleep apnea and control subjects to nonactivated human umbilical vein endothelial cells in culture||

Adhesion Index of γδ T Cells (n)*

Adhesion Index of γδ T Cells (n)*

Without anti–TNF-α
With anti–TNF-α
Percentage of γδ
 T Cells Bearing
 L-Selectin (n)
With Isotypic
 Control Cells§
 anti-CD62E /CD62P||
OSA6.1 ± 0.9 (9)2.9 ± 0.7 (9)65.0 ± 11.7 (11)6.1 ± 1.1 (5)2.9 ± 0.2 (5)
OSA relative, %100 ± 14.747.5 ± 11.4100 ± 18.047.5 ± 3.3
p Value0.00020.006
Control cells3.6 ± 0.9(6)**3.6 ± 0.7 (6)49.6 ± 11.0 (8)**3.4 ± 0.5 (5)**2.6 ± 0.2 (5)
Control cells
   relative, %100 ± 25.0100 ± 19.4100 ± 14.776.5 ± 5.9
p Value



*Magnet-separated γδ T cells from the same donors were used in parallel experiments.

Adhesion assays were performed without anti-human TNF-α.

Adhesion assays were performed with 500 ng/ml of anti-human TNF-α.

§HUVEC confluent monolayers were pretreated for 1 h at 37°C with isotypic control antibodies.

||HUVEC confluent monolayers were pretreated for 1 h at 37°C with anti-CD62E/CD62P antibodies.

p Values for each group with and without the appropriate antibody.

**Statistically significant control cells vs. OSA, p < 0.01.

Definition of abbreviations: HUVEC = human umbilical vein endothelial cell; NS = not significant; OSA = obstructive sleep apnea; TNF = tumor necrosis factor.

Adjusted means ± SD are presented. The relative percentage of patients with OSA and control subjects depicts the values obtained in the presence of antibodies compared with treatments without antibodies.


L-selectin (CD62L) expression was also significantly higher in OSA γδ T cells than in control cells (Table 2, p < 0.01). Thus, pretreating HUVECs with monoclonal antibodies against CD62E/CD62P, which block binding through selectin receptors, significantly inhibited the adhesion index of OSA γδ T cells by 52.5% (p < 0.006). In control cells the adhesion index was lowered only by 23.5% (p < 0.01).

In order to determine whether the increased avidity between OSA γδ T cells/HUVECs was also affected by the presence of NK receptors (CD56 and CD16), in parallel experiments from the same donors we used both magnet-separated γδ T cells and γδ T cells depleted from the CD56- and CD16-bearing cells by the RosetteSep antibody cocktail (see methods). Depletion of cells bearing CD56 and CD16 NK receptors from the γδ T cell population resulted in a decrease of about 22% in the adhesion index from 6.3 ± 1.2 to 4.9 ± 1.1 in 5/5 of patients with OSA. A similar trend was observed in control subjects (data not shown).

Collectively, the avidity of OSA γδ T cells to nonactivated HUVECs in culture was increased due to the involvement of TNF-α and the presence of L-selectin receptors and to a lesser extent due to the NK receptors CD56 and CD16.

Response of γδ T Cells to IL-2/IL-7 Treatment

To further characterize the γδ T cells of patients with OSA, we examined the response of lymphocytes (1,000 cells/well) cultured on HUVECs in the presence of IL-2 (20 U/ml) and IL-7 (0.5 ng/ml). IL-7 is an essential cytokine for early γδ T cell development (27). We found that T lymphocytes of patients with OSA had a higher response to activation in vitro by the IL-2/IL-7 cocktail as compared with control cells. During 4 to 5 days of culturing with cytokines, high blastogenesis and heavy proliferation were noted in OSA lymphocyte cocultures as compared with control cocultures (data not shown). The number of clones stained with fluorescein isothiocyanate–conjugated monoclonal antibodies against γδ T cells, which developed after 4 to 5 days of coculturing with HUVEC monolayers, was also significantly higher in patients with OSA as compared with control subjects (5.7 ± 1.5 clones, n = 7 vs. 1.5 ± 1.6 clones, n = 4, respectively; p = 0.006).

Cytotoxicity against Endothelial Cells: Involvement of TNF-α and NK Receptors

The cytotoxicity of γδ T cells against HUVECs was studied in cocultures at the ratio effector/target of 5:1 for 2 days. Cytotoxicity of OSA γδ T cells against HUVECs was 2.5-fold higher than that for control subjects (Figure 4

, p = 0.0001). Moreover, the signs of DNA fragmentation in HUVECs, as determined by propidium iodide labeling, were evident at sites of adhesion of γδ T cells of patients with OSA within 24 hour of lymphocyte addition to cultures (data not shown). These cytotoxic effects could largely be attributed to TNF-α because adding anti–TNF-α abolished the killing of HUVECs (p = 0.003) by γδ T cells (Figure 4). As in previous experiments, the individual data are plotted. Three out of 8 control subjects were current smokers, and their cytotoxicity values were slightly higher than that of nonsmoker control subjects. In OSA, in 5/14 current smokers the distribution of cytotoxicity values was similar to that of nonsmokers. Thus, smoking did not seem to affect the data in OSA.

As in the case of the adhesion assay, removal of γδ T cells bearing CD16 and CD56 receptors did not affect the results, thereby indicating that the contribution of CD16 and CD56 receptors was minor (in OSA γδ T cells, the percentage of specific 51Cr release was lowered from 45.9 ± 3.5% to 39.3 ± 6.7% in 4/4 cases).

To our knowledge, this study demonstrates for the first time the potential significance of γδ T cells to endothelial cell injury in OSA. Cumulative evidence indicates that OSA is associated with cardiovascular disease morbidity and mortality (13). Moreover, prominent mechanisms in cardiovascular disease such as endothelial dysfunction, the earliest manifestation of atherosclerosis, were shown to be accentuated in patients with sleep apnea (28, 29). Two major approaches are currently under intense investigation: augmented sympathetic activation associated with sleep fragmentation and increased oxidative stress due to hypoxia/reoxygenation injury and possibly hypercapnia. Both have consequences to cardiovascular function.

The intermittent hypoxia experienced by patients with OSA is somewhat analogous to the hypoxia/reperfusion–reoxygenation phenomenon (30). This has been implicated in microvascular dysfunction due to increased oxidative stress and active participation of circulating inflammatory cells (31). The contribution of each of the circulating inflammatory cells, and their specific roles, however, are poorly understood in the setting of the hypoxia/reoxygenation phenomenon. Yet, in recent years a rapid advance has been made in the understanding of inflammatory–immunologic mechanisms that govern atherogenesis. Currently available data have shown that immunologic activation is an early step in this process (32).

Atherosclerosis consists of a cascade of events leading to the formation of focal lesions in the arterial intima, which are characterized by cholesterol deposition, fibrosis, and inflammation (33). These lesions begin as local infiltrates of monocyte-derived macrophages, T lymphocytes, lipoproteins, adhesion molecules, and cytokines (34, 35). The cell populations recruited to atherosclerotic lesions are heterogeneous, and besides macrophages include also αβ T cells expressing CD4 and CD8 receptors and γδ T cells (16, 32, 36). A large proportion of them express an activated phenotype (32, 36, 37). This activation of macrophages and T cells that leads to cytokine production and induces an inflammatory state promotes endothelial and vascular smooth muscle cell activation as well (38). Moreover, T cell activation in atherosclerotic plaques was suggested to injure endothelial cells via release of inflammatory cytokines and T cell cytotoxicity (16, 32, 36).

Because we have previously shown that monocytes of patients with OSA were activated (4), we sought to expand on these findings, and thus demonstrated that γδ T lymphocytes of patients with OSA, express an activated phenotype as well. This was manifested by altered cytokine balance and increased adhesion and cytotoxicity toward endothelial cells in culture.

Determination of the surface expression of γδ T cell receptors in patients with OSA and control subjects in whole blood revealed no differences between the groups. Hence, despite the similar percentage of circulating γδ T cells, the number of γδ T cell clones that were detected in lymphocytes/HUVECs cocultures in the presence of IL-2/IL-7 cocktail was higher in patients with OSA. Because IL-7 is essential for the proliferation, survival, and differentiation of γδ T cells from T cell precursors (27), it suggests an increase in γδ progenitor cells of patients with OSA. Moreover, this increased amount of γδ T cells observed at the sites of adhesion to endothelial cells in culture in vitro may indicate increased accumulation of γδ T cells on the endothelium lining the blood vessels in patients with OSA.

On the basis of the data that a large proportion of circulating γδ T cells expresses different types of NK receptors, such that regulate T cell function (39), their expression was specifically analyzed on γδ T cells of these patients. Paradoxically, the percentage of γδ T cells of patients with OSA, which expressed the inhibitory NKB1 receptors, was significantly higher than that of control subjects (Figure 1), even though the cytotoxicity of OSA γδ T cells against HUVECs was increased (Figure 4). This type of NK receptor participates in the inhibition of cell-mediated cytotoxicity and cytokine secretion. Because the cytotoxic function is a result of a balance between activating and inhibiting signals, which are delivered by the corresponding receptors (40), it is conceivable that increased inhibitory activity of γδ T cells in OSA provides a compensatory mechanism to avoid endothelial injury. On the other hand, the percentage of γδ T cells bearing CD56 and CD16 receptors did not vary between patients with OSA and control subjects. Moreover, the values were similar to values reported for healthy donors (41). The participation of the CD56 receptor in direct cytotoxicity, however, was described for other lymphocyte subpopulations (42). In our study, despite the similar expression of CD56 and CD16 between γδ T cells of patients with OSA and control subjects, patients' γδ T cells exhibited a 2.5-fold higher cytotoxicity against HUVECs. In addition, depleting CD56- and CD16-bearing cells from γδ T cells resulted only in a minor decrease in their cytotoxicity against HUVECs. Thus, although CD56 and CD16 receptors could be involved in the cytotoxicity induced by OSA γδ T cells, their contribution to damage HUVECs in culture was found to be limited.

The currently available data indicate that reperfusion/reoxygenation that follows hypoxic periods (43), activates a variety of cells including endothelial cells and leukocytes, thus propagating inflammatory processes (43). As these cells become activated either via hypoxia/reoxygenation directly or via inflammatory cytokines such as TNF-α, they express adhesion molecules that contribute to increased endothelial cells/inflammatory cells interactions and increased adhesion to vascular walls, eventually initiating atherogenic processes (43, 44). Moreover, the classic proinflammatory cytokines TNF-α, IL-6, and IL-8 are known to be regulated by oxygen tension and free radicals, which are prominent mechanisms in hypoxia/reoxygenation injury (45). In accord with this course of events, we found an increase in the percentage of OSA γδ T cells expressing TNF-α and IL-8 and a decrease with respect to the antiinflammatory cytokine IL-10, altering the balance between proinflammatory and antiinflammatory cytokines in OSA, and by that further increasing the deleterious effects due to overactivation of these cells (Table 1, Figure 2). Consequently, the adhesion of OSA γδ T cells to nonactivated HUVECs in culture was found to be twofold higher than that in control cells. This could be largely attributed to the increases observed in the percentage of γδ T cells expressing L-selectin receptors and TNF-α as verified by the pretreatment of HUVECs with specific monoclonal antibodies. Using CD62E/CD62P antibodies, which blocked the binding through selectin receptors, significantly inhibited OSA γδ T cells adhesion index by about 50% (Table 2). Similarly, neutralization of the endogenous TNF-α by adding anti TNF-α antibodies abolished both the adhesion and the cytotoxicity against HUVEC monolayers by OSA γδ T cells (Table 2, Figure 4). These data provide evidence for a potentially important pathway in which augmented TNF-α secretion at sites of γδ T cell adhesion may play a major role in injuring the endothelium. Moreover, our experiments were conducted with nonactivated HUVECs, yet extremely strong cytotoxicity was noted when cocultured with OSA γδ T cells. The possibility that in vivo–activated endothelium, due to intermittent hypoxia and increased proinflammatory cytokine production, may result in exacerbated adhesiveness, should be considered.

TNF-α is now recognized as a critical cytokine orchestrating differentiation and proliferation as well as the ability to induce cell death. As a proinflammatory cytokine, TNF-α was implicated in several diseases with joint and tissue destruction (46). Because of its ability to cause cell necrosis, new blood vessel formation, and increased thrombogenicity TNF-α was implicated in atheroma formation (47) and in cytotoxicity and thrombogenicity against endothelial cells (48). Of note, plasma levels of TNF-α were found to be elevated in patients with sleep apnea as compared with control subjects and patients with excessive daytime sleepiness but without sleep apnea (7, 8). IL-8, another proinflammatory cytokine, which we found to be expressed in a higher percentage in OSA γδ T cells relative to control cells, was also shown to be increased in the circulation of patients with OSA (9). Because this cytokine exerts chemotactic effects on neutrophils (49), lymphocytes (50), and activated NK cells (51), increased IL-8 production in OSA could possibly result in increased accumulation of neutrophils and lymphocytes at sites of γδ T cell adhesion to the endothelium. On the other hand, IL-10, which was found to be inversely correlated with the expression of TNF-α and IL-8, has emerged in recent years as an important regulator of the immune and inflammatory systems (52, 53). It inhibits the production of proinflammatory cytokines (IL-1β, TNF-α, IL-6, IL-8, and IL-12) and the release of free radicals by monocytes and/or macrophages (54), and by that serves as a protective mechanism in the progression of atherosclerotic lesion formation and its stability (55, 52). Consequently, patients with unstable angina had significantly lower serum IL-10 concentrations than patients with chronic stable angina (56). Furthermore, IL-10 was shown to protect ischemic and reperfused myocardium through the suppression of neutrophil recruitment, and a decrease in TNF-α, and in intracellular adhesion molecule-1 (57). In addition, endothelial dysfunction and superoxide production were increased during inflammation in IL-10–deficient mice, thus providing direct evidence that IL-10 protects endothelial function after acute inflammatory stimulus by limiting local increases in superoxide (58).

Intracellular signals regulating cytokine production in γδ T cells are poorly understood. In vitro activation of human γδ T cells by nonpeptidic ligands rapidly induces a massive production of TNF-α (59). However, a more likely possibility in OSA is that cytokine production may be modulated by changes in the surrounding oxygen tension because inflammatory cytokines are known to be regulated by oxidative stress and free radicals, via activation of redox-sensitive transcription factors (45, 54). For instance, hypoxia in vitro was shown to enhance IL-2, IL-4, IL-6, and IFN-γ production and to inhibit IL-10 release from resting or stimulated peripheral blood mononuclear cells (60). IL-8 messenger RNA was specifically induced by hypoxia in dermal fibroblasts (61). Moreover, exposure of endothelial cells to anoxia and anoxia/reoxygenation resulted in increased adhesiveness of endothelial cells to T lymphocytes that was followed by increased TNF-α production and increased adhesion of neutrophils via specific adhesion molecules (62, 63). This implicates redox imbalance in the molecular mechanisms of neutrophil/endothelial cell adhesion (63). It is therefore reasonable to assume that the intermittent hypoxia that patients with OSA experience nightly, which results in increased TNF-α and decreased IL-10 production from activated γδ T cells and possibly from other sources, upregulates the expression of adhesion molecules on endothelial cells and thus promotes adhesion of monocytes, lymphocytes, and possibly neutrophils.

Whereas redox imbalance is one possible explanation (4, 5, 63), we cannot firmly conclude from this study if altered γδ T cell function solely results from the direct effects of hypoxia/reperfusion, as seen for some adhesion molecules expressed in OSA monocytes (4), or from a combined effect of hypoxia/reperfusion and hypercapnia (64), and/or sleep fragmentation–related sympathetic activation (65). Although hypercapnia is a prominent feature of OSA, very few studies have addressed this question at the cellular and molecular level. On the other hand, regarding sympathetic activation, cumulative evidence indicates that the sympathetic nervous system innervates all lymphoid organs, and that catecholamines, the end products of sympathetic activation, modulate several immune parameters (66). Therefore, the sympathetic activation and the increased catecholamine (norepinephrine) release observed in patients with OSA (67) could affect lymphocyte traffic, circulation, and proliferation, and modulate cytokine production and functional activities of lymphoid cells as well (66, 68). Additional studies merit this question.

About 21% of the patients and 29% of the control subjects studied were current smokers. Although this could be a potential limitation in the study, overall, smoking did not seem to dramatically affect circulating γδ T functions in patients with OSA or control subjects as depicted by the individual data. However, some parameters in smoking controls were on the high side, but they did not mask the effects of OSA. Moreover, an extensive study on lymphocyte subsets of smokers clearly stated that neither the percentage nor the absolute values of peripheral blood γδ T cells were affected (69). On the other hand, significant increases in γδ T cell numbers from the bronchial wall sections of smokers were found (70). In addition, blood was collected on awakening from sleep, before eating, or before smoking, and subjects had at least 10 hours of smoke-free conditions before blood collection. We should also note that many of the cellular functions we investigated are relatively quick to respond (overnight). For instance, a single nasal continuous positive air pressure night treatment lowered the expression of adhesion molecules, decreased adhesion index and basal free radical production by OSA monocytes, whereas omitting nCPAP for a single night from otherwise treated patients, immediately restored the values to patients' levels (4).

In conclusion, γδ T cells of patients with OSA express increased NKB1 inhibitory receptors, increased content of proinflammatory cytokines, and a decrease in antiinflammatory cytokine content. Consequently, avidity and cytotoxicity against endothelial cells in culture are increased. Employing neutralizing antibodies against E/P-selectin attenuates the adhesion response. Similarly, anti– TNF-α attenuates the avidity and abolishes the cytotoxicity against endothelial cells in culture, all specifically implicating the involvement of γδ T cells in initiating or accentuating endothelial cell damage in OSA.

The help and support of the staff of the Technion Sleep Lab are gratefully acknowledged. Z. Tsabari and F. Barbara provided invaluable technical help, and P. Herer provided statistical consultation.

1. Lavie P, Herer P, Hoffstein V. Obstructive sleep apnoea syndrome as a risk factor for hypertension: population study. BMJ 2000;320:479–482.
2. Shahar E, Whitney CW, Redline S, Lee ET, Newman AB, Javier Nieto F, O'Connor GT, Boland LL, Schwartz JE, Samet JM for the Sleep Heart Health Study Research Group. Sleep-disordered breathing and cardiovascular disease cross-sectional results of the Sleep Heart Health Study. Am J Respir Crit Care Med 2001;163:19–25.
3. Malhotra A, White D. Obstructive sleep apnea. Lancet 2002;360:237–245.
4. Dyugovskaya L, Lavie P, Lavie L. Increased adhesion molecules expression and production of reactive oxygen species in leukocytes of sleep apnea patients. Am J Respir Crit Care Med 2002;165:934–939.
5. Lavie L. Obstructive sleep apnoea syndrome: an oxidative stress disorder. Sleep Med Rev 2003;7:35–51.
6. Monaco C, Andreakos E, Young S, Feldmann M, Paleolog E. T cell mediated signaling to vascular endothelium: induction of cytokines, chemokines, and tissue factor. J Leukoc Biol 2002;71:659–668.
7. Vgontzas AN, Papanicolaou DA, Bixler EO, Kales A, Tyson K, Chrousos GP. Elevation of plasma cytokines in disorders of excessive daytime sleepiness: role of sleep disturbances and obesity. J Clin Endocrinol Metab 1997;82:1313–1316.
8. Vgontzas AN, Papanicolaou DA, Bixler EO, Hopper K, Lotsikas A, Lin HM, Kales A, Chrousos GP. Sleep apnea and daytime sleepiness and fatigue: relation to visceral obesity, insulin resistance and hypercytokinemia. J Clin Endocrinol Metab 2000;85:1151–1158.
9. Ohga E, Tomita T, Wada H, Yamamoto H, Nagase T, Ouchi Y. Effects of obstructive sleep apnea on circulating ICAM-1, IL-8, and MCP-1. J Appl Physiol 2003;94:179–184.
10. Hayday AC. γδ Cells: a right time and a right place for a conserved third way of protection. Annu Rev Immunol 2000;18:975–1026.
11. Bluestone JA, Khattri R, Sciammas R, Sperling AI. TCR gamma delta cells: a specialized T cell subset in the immune system. Annu Rev Cell Dev Biol 1995;11:307–353.
12. Crowley MP, Reich Z, Mavaddat N, Altman JD, Chein Y-H. The recognition of the non-classical major histocompatibility complex (MHC) class 1 molecules, T10, by the γδ T-cell, G8. J Exp Med 1997;185:1223–1230.
13. Young AJ, Marston WL, Dudler L. Subset-specific regulation of the lymphatic exit of recirculating lymphocytes in vivo. J Immunol 2000;165: 3168–3174.
14. Carding SR, Egan PJ. Gammadelta T-cells: functional plasticity and heterogeneity. Nat Rev Immunol 2002;2:336–345.
15. Song L, Leung C, Schindler C. Lymphocytes are important in early atherosclerosis. J Clin Invest 2001;108:251–259.
16. Emeson EE, Robetson AL. T lymphocytes in aortic and coronary intimas: their potential role in atherogenesis. J Pathol 1988;130:369–376.
17. Kleindienst R, Xu Q, Willeit J, Waldenberger FR, Weimann S, Wick G. Immunology of atherosclerosis: demonstration of heat shock protein 60 expression and T lymphocytes bearing alpha/beta or gamma/delta receptor in human atherosclerosic lesions. Am J Pathol 1993;142:1927–1937.
18. Wick G, Xu Q. Atherosclerosis: an autoimmune disease. Exp Gerontol 1999;34:559–566.
19. Galea P, Brezinschek R, Lipsky PE, Oppenheimer-Marks N. Phenotyic characterization of CD4-/alpha beta TCR+ and gamma delta TCR+ T cells with a transendothelial migratory capacity. J Immunol 1994;153: 529–542.
20. Kahaleh MB, Fan PS, Otsuka T. Gamma/delta receptor bearing T cells in scleroderma: enhanced interaction with vascular endothelial cells in vitro. Clin Immunol 1999;91:188–195.
21. Mohagheghpour N, Bermuder LE, Khajavi S, Rivas A. The VLA-4/VCAM-1 molecules participate in gamma delta cell interaction with endothelial cells. Cell Immunol 1992;143:170–182.
22. Kretowski A, Mysliwiec J, Szelachowska M, Turowski D, Wysocka J, Kowalska I, Kinalska I. Gammadelta T cells alterations in the peripheral blood of high risk diabetes type 1 subjects with subclinical pancreatic B cell impairment. Immunol Lett 1999;68:289–293.
23. Bank I, Chohen L, Mouallem M, Ferfel Z, Grossman E, Ben-Nun A. γδ T cell subsets in patients with arthritis and chronic neutropenia. Ann Rheum Dis 2002;61:438–443.
24. Heng MK, Heng MC. Heat shock protein 65 and activated gamma/delta T cells in injured arteries. Lancet 1994;344:921–923.
25. Dyugovskaya L, Bashenko Y, Hirsh M, Krausz M. Immune potential of lymph node-derived lymphocytes in uncontrolled hemorrhagic shock. Shock 2001;15:118–123.
26. Nicoletti I, Migliorati G, Pagliacci MC, Grignani F, Riccardu C. A rapid simple method for measuring thymocyte apoptosis by propidium iodide staining and flow cytometry. J Immunol Methods 1991;139:271–279.
27. Ikuta K, Lee H-C, Ye S-K. Role of the IL-7 receptor in gamma-delta T cell development. Chem Immunol 2001;79:29–42.
28. Kato M, Roberts-Thomson P, Phillips BG, Haynes WG, Winnicki M, Accurso V, Somers VK. Impairment of endothelium-dependent vasodilation of resistance vessels in patients with obstructive sleep apnea. Circulation 2000;102:2607–2610.
29. Kraiczi H, Caidahl K, Samuelsson A, Peker Y, Hedner J. Impairment of vascular endothelial function and left ventricular filling: association with the severity of apnea-induced hypoxemia during sleep. Chest 2001; 119:1085–1091.
30. Dean RT, Wilcox I. Possible atherogenic effects of hypoxia during obstructive sleep apnea: (supplement: 2. natural history of sleep apnea). Sleep 1993;16:S15–S22.
31. Shigematsu S, Wolf RE, Granger DN. T lymphocytes modulate the microvascular and inflammatory response to intestinal ischamia-reperfusion. Microcirculation 2002;9:99–109.
32. Hansson GK. Immune mechanisms in atherosclerosis. Arterioscler Thromb Vasc Biol 2001;21:1876–1890.
33. Ross R. Atherosclerosis: an inflammatory disease. N Engl J Med 1999;340: 115–126.
34. Steinberg D, Parthasarathy S, Carew TE, Khoo JC, Witzum JL. Beyond cholesterol: modifications of low-density lipoprotein that increase its atherogenicity. N Engl J Med 1989;320:915–924.
35. Stemme S, Faber B, Holm J, Wiklund O, Witztum JL, Hansson GK. T lymphocytes from human atherosclerotic plaques recognize oxidized low density lipoprotein. Proc Natl Acad Sci USA 1995;92:3893–3897.
36. Stemme S, Holm J, Hansson GK. T lymphocytes in human atherosclerotic plaques are memory cells expressing CD45RO and the integrin VLA-1. Arterioscler Thromb 1992;12:206–211.
37. Van der Wal AC, Das PK, van de Berg DB, van der Loos CM, Becker AE. Atherosclerotic lesions in humans: in situ immunophenotypic analysis suggesting an immune mediated response. Lab Invest 1989;61: 166–170.
38. Pober JS, Collins T, Gimbrone MA Jr, Cotran RS, Gitlin JD, Fiers W, Clayberger C, Krensky AM, Burakoff SJ, Reiss CS. Lymphocytes recognize human vascular endothelial and dermal fibroblast Ia antigens induced by recombinant immune interferon. Nature 1983;305: 726–729.
39. Battistini L, Borsellino G, Sawicki G, Poccia F, Salvetti M, Ristori G, Brosnan CF. Phenotypic and cytokine analysis of human peripheral blood γδ T cells expressing NK cell receptors. J Immunol 1997;159: 3723–3730.
40. Solana R, Mariani E. NK and NK/T cells in human senescence. Vaccine 2000;18:1613–1620.
41. Zheng BJ, Ng SP, Chua DT, Sham JS, Kwong DL, Lam CK, Ng MH. Peripheral gamma delta T cell deficit in nasopharyngeal carcinoma. Int J Cancer 2002;99:213–217.
42. Pittet MJ, Speiser DE, Valmori D, Cerottini J-C, Romero P. Cutting edge: cytolytic effector function in human circulating CD8+ T cells closely correlates with CD56 surface expression. J Immunol 2000;164: 1148–1152.
43. Grisham MB, Granger DN, Lefer DJ. Modulation of leukocyte-endothelial interactions by reactive metabolites of oxygen and nitrogen: relevance to ischemic heart disease. Free Radic Biol Med 1998;25:404–433.
44. Walzog B, Gaehtgens P. Adhesion molecules: the path to a new understanding of acute inflammation. News Physiol Sci 2000;15:107–113.
45. Haddad JJ. Pharmaco-redox regulation of cytokine-related pathways: from receptor signaling to pharmacogenomics. Free Radic Biol Med 2002;33:907–926.
46. Kollias G, Douni E, Kassiotis G, Kontoyiannis D. On the role of tumor necrotic factor and receptors in models multiorgan failure, rheumatoid arthritis, multiple sclerosis and inflammatory bowel disease. Immunol Rev 1999;169:175–194.
47. Barath P, Fishbein MC, Cao G, Berenson L, Helfant RH, Forrester JS. Detection and localization of tumor necrotic factor in human atheroma. Am J Cardiol 1990;65:297–302.
48. Bevilacqua MP, Pober JS, Majeau GR, Fiers W, Cotran RS, Gimbrone MA. Recombinant tumor necrotic factor induces procoagulant activity in cultured human vascular endothelium: characterization and comparison with the action of interleukin-1. Proc Natl Acad Sci USA 1986;83: 4533–4537.
49. Van Damme J, Van Beeumen J, Opdenakker G, Billau A. A novel, NH2-terminal sequence-characterized human monokine possessing neutrophil chemotactic, skin-reactive, and granulocytosis-promoting activity. J Exp Med 1988;167:1364–1376.
50. Koch AE, Polverini PJ, Kunkel SL, Harlow LA, DiPietro LA, Elner VM, Elner SG, Strieter RM. Interleukin-8 as a macrophage-derived mediator of angiogenesis. Science 1992;258:1798–1801.
51. Sebok K, Woodside D, al-Aoukaty A, Ho AD, Gluck S, Maghazachi AA. IL-8 induces the locomotion of human IL-2-activated natural killer cells: involvement of guanine nucleotide binding (Go) protein. J Immunol 1993;150:1524–1534.
52. Terkeltaub RA. IL-10: an “immunologic scalpel” for atherosclerosis? Arterioscler Thromb Vasc Biol 1999;19:2823–2825.
53. Mallat Z, Besnard S, Duriez M, Deleuze V, Emmanuel F, Bureau MF, Soubrier F, Esposito B, Duez H, Fievet C, et al. Protective role of interleukin-10 in atherosclerosis. Circ Res 1999;85:17–24.
54. Haddad JJ, Fahlman CS. Redox- and oxidant-mediated regulation of interleukin-10: an inflammatory, antioxidant cytokine? Biochem Biophys Res Commun 2002;297:163–176.
55. Pinderski Oslund LJ, Hedrick CC, Olvera T, Hagenbaugh A, Territo M, Berliner JA, Fyfe AI. Interleukin-10 blocks atherosclerotic events in vitro and in vivo. Arterioscler Thromb Vasc Biol 1999;19:2847–2853.
56. Smith DA, Irving SD, Sheldon J, Cole D, Kaski JC. Serum level of the anti-inflammatory cytokine interleukine-10 are decreased in patients with unstable angina. Circulation 2001;104:746–749.
57. Yang Z, Zingarelli B, Szabo C. Crucial role of endogenous interleukin-10 production in myocardial ischemia/reperfusion injury. Circulation 2000;101:1019–1026.
58. Gunnett CA, Heistad DD, Berg DJ, Faraci FM. IL-10 deficiency increases superoxide and endothelial dysfunction during inflammation. Am J Physiol Heart Circ Physiol 2000;299:H1555–H1562.
59. Lafont V, Liautard J, Liautard JP, Favero J. Production of TNF-alpha by human V gamma 9V delta 2 T cells via engagement of Fc gammaRIIIA, the low affinity type 3 receptor for the Fc portion of IgG, expressed upon TCR activation by nonpeptidic antigen. J Immunol 2001;166: 7190–7199.
60. Naldini A, Carraro F, Sivestri S, Bocci V. Hypoxia affects cytokine production and proliferation response by human peripheral mononuclear cells. J Cell Physiol 1997;173:335–342.
61. Galindo M, Santiago B, Alcami J, Rivero M, Martin-Serrano J, Pablos JL. Hypoxia induces expression of the chemokines monocyte chemoattractant protein-1 (MCP-1) and IL-8 in human dermal fibroblasts. Clin Exp Immunol 2001;23:36–41.
62. Kokura S, Wolf RE, Yoshikawa T, Granger DN, Aw TY. T-lymphocyte-derived tumor necrosis factor exacerbates anoxia-reoxygenationinduced neutrophil-endothelial cell adhesion. Circ Res 2000;86:205–213.
63. Kokura S, Yoshida N, Yoshikawa T. Anoxia/reoxygenation-induced leukocyte-endothelial cell interactions. Free Radic Biol Med 2002;33:427–432.
64. Chen L, Vicaut E, Sercombe R. Polymorphonuclear leukocyte activation induces cerebral hypoperfusion in rats in the absence of previous ischemia-reperfusion. Neurosci Lett 2002;331:203–207.
65. Lofaso F, Coste A, Gilain L, Harf A, Guilleminault C, Goldenberg F. Sleep fragmentation as a risk factor for hypertension in middle aged non-apneic snorers. Chest 1996;109:896–900.
66. Elenkov IJ, Wilder RL, Chrousos GP, Vizi ES. The sympathetic nerve: an integrative interference between two supersystems: the brain and the immune system. Pharmacol Rev 2000;52:595–638.
67. Peled N, Greenberg A, Pillar G, Zinder O, Levi N, Lavie P. Contributions of hypoxia and respiratory disturbance index to sympathetic activation and blood pressure in obstructive sleep apnea syndrome. Am J Hypertens 1998;11:1284–1289.
68. Elenkov IJ, Chrousos GP, Wilder RL. Neuroendocrine regulation of IL-12 and TNF-α/IL-10 balance: clinical implications. Ann N Y Acad Sci 2000;917:94–105.
69. Schaberg T, Theilacker C, Nitschke OT, Lode H. Lymphocyte subsets in peripheral blood and smoking habits. Lung 1997;175:387–394.
70. Richmond I, Pritchard GE, Ashcroft T, Corris PA, Walters EH. Distribution of gamma delta T cells in the bronchial tree of smokers and non-smokers. J Clin Pathol 1993;46:926–930.
Correspondence and requests for reprints should be addressed to Lena Lavie, Ph.D., Unit of Anatomy and Cell Biology, The Bruce Rappaport Faculty of Medicine, Technion, POB 9649, 31096 Haifa, Israel. E-mail:


No related items
American Journal of Respiratory and Critical Care Medicine

Click to see any corrections or updates and to confirm this is the authentic version of record